# 9.2 Comparing Logistic, Poisson, and Survival Analysis

## Метаданные

- **Канал:** MarinStatsLectures-R Programming & Statistics
- **YouTube:** https://www.youtube.com/watch?v=o7hGttZhLlc

## Содержание

### [0:00](https://www.youtube.com/watch?v=o7hGttZhLlc) Segment 1 (00:00 - 05:00)

before we get into the details of poisson regression what i want to do is spend a little bit of time building some connections between logistic regression poisson regression and survival analysis they're all very much related and the model that we might choose to use depends partly on the question the research question that we're asking and trying to answer so what i want to do is just quickly compare and contrast the three of these and show the way we might use the models um in the context of i'm nsa let's imagine that we're interested in studying something about an outcome variable of getting the flu okay so if someone gets a flu in a particular season so just a reminder logistic regression what we do there is we use it to model if an event occurs right so does it occur yes or no so in this context when we're looking at the flu over a particular season we can think of asking a question does an individual get the flu yes or no right and now one important thing with logistic regression we need to have some endpoint so we're talking about um does you know well the example we used earlier in this course is a baby low birth weight yes or no we observe that end point remember saying to someone to get the flu that's too much of an open-ended question of what do we mean by that within the next week the next month in a year over their entire lifetime because we need to specify some endpoint does the individual get the flu and let's say in the 2019-2012 20 winter so defining some end point where we look at in this particular winter season will someone get the flu yes or no so that's the previous winter poisson regression i think here we're modeling the rate at which the event occurs and again so we can model counts or rates so in the same context we might try to model something like how many times does someone get the flu and again let's say in the 2019 2020 winter some year that's already the end points are done and we can count the number of times someone got the flu over this season okay so you can see they're very much related questions does someone get the flu yes or no how often did they get the flu survival analysis what we're going to model there is we're the time until the event occurs so in this same context we could think of asking a question like how long until someone gets the flu and again here you have some start point defined let's consider measuring from the start of the 2019 winter okay so from the first day of the 2019 winter once that starts how much time goes by until someone gets the flu so the point of me doing this was i wanted to show you that these three models are all very much related it probably depends on the type of question we want to ask does the event happen how often does it how long till it happens and we've grouped together in this course poisson regression and survival analysis because these two are also very much related these are both looking at events that occur over time right we're looking at if we define a period of time right the winter how often does the flu

### [5:00](https://www.youtube.com/watch?v=o7hGttZhLlc&t=300s) Segment 2 (05:00 - 05:00)

occur or this one we're asking specifically about the time right how much time goes by until someone gets the food so what we're going to do now is move into building some of the underlying theory for poisson regression and survival analysis we'll talk about what the poisson process is and then we'll focus on poisson regression for a few weeks and then we'll move to survival analysis after that stick around guys there's more to see and please stay safe

---
*Источник: https://ekstraktznaniy.ru/video/44717*